Release · v39.0 · current

Download DawaImg for your platform.

Free. Offline. No accounts, no accounts, no license, no activation. Pick your OS below — the button is already highlighted for your system.

Windows · Installer (.msi)
v39.0.0~236 MBx64
Download for Windows Release notes & checksum
macOS · Disk image (.dmg)
v39.0~366 MBUniversal
Download for macOS Release notes & checksum

Every download CTA is measured as a GA4 download_click event (platform + location) and mirrored to a private local activity log. See “Your download activity” below.

System requirements

Runs on the machine you already have.

DawaImg is a native desktop app. No browser, no VM, no cloud. Requirements are modest; performance scales with local CPU and RAM.

Windows

  • Windows 10 or 11 (64-bit)
  • 8 GB RAM (16 GB+ for large datasets)
  • ~6 GB free disk (install + project cache)
  • No internet connection required

macOS

  • macOS 12 Monterey or later
  • 8 GB RAM (16 GB+ for large datasets)
  • ~8 GB free disk (install + project cache)
  • Apple Silicon & Intel supported
Install in under a minute

Three steps, no wizarding required.

Windows

  1. Run DawaImg-39.0.0.msi.
  2. Follow the Installer — pick a folder, next, done.
  3. Launch DawaImg and create your first project.
Download for Windows

macOS

  1. Open DawaImg-39.dmg and drag DawaImg into Applications.
  2. On first launch, allow local folder access when prompted.
  3. Launch DawaImg and create your first project.
Download for macOS
What’s in the box

Everything, in one installer. Nothing hidden.

  • Face detection & extraction
  • People clustering
  • Similarity search
  • Face blurring / redaction
  • Provenance record chain
  • Reports & exports
  • Detection labs
  • Project manager
Release · v39.0

Verification & safety

DawaImg installers are signed. No “ring home” on first launch. No telemetry hooks, no update pings.

Windows · DawaImg-39.0.0.msi

247,324,680 bytes · ≈ 236 MiB

macOS · DawaImg-39.dmg

383,300,763 bytes · ≈ 366 MiB

SHA-256 checksums are published on the GitHub releases page. Verify after download.

Changelog

v39.0 highlights

Added

  • Detection lab for robustness evaluation
  • Face-mask design & failure-mode tooling
  • Minor-redaction export option
  • Time-windowed video face counts

Improved

  • 2.4× faster frame extraction on Apple Silicon
  • Reduced memory on multi-million-file projects
  • Cleaner provenance graph in reports
  • Windows MSI: per-machine and per-user install
Public research bundles

Try DawaImg on real data.

Download project bundles built from publicly released sources, open them in DawaImg, and explore the exact workflow used in our case study.

DatasetAccessScaleDate
Public document release — House Oversight
Extracted faces from a publicly released document archive. Extracted faces are unlabeled.		 Request access
Read report		 ~1,000,000 files · ~6 h local Feb 4, 2026
Public document release — DOJ disclosures
Extracted faces from publicly released DOJ disclosure documents.		 Request access ~100,000 documents Feb 4, 2026
Minnesota public media archive
Archive-grade organization of publicly available image & video material.		 Contribute on GitHub Large-scale public media Dec 2025

How to import projects

  1. Download the .zip bundle.
  2. Unzip the file into a folder.
  3. Place the folder inside ~/DawaImg/projects/<Project>/ (or %USERPROFILE%\DawaImg\projects\<Project>\ on Windows).
  4. Launch DawaImg — the project appears on the Home screen.

Ready to start on your own machine?

Free. Offline. No accounts. No cloud.

Download for Windows Download for macOS